9.4.0. Survival Variable Selection
Enter Durations: One column containing the duration of each case is selected by clicking on [Time].
Enter Begin and End Times: Select one column as the starting time by clicking on [Begin] and another column as the final time by clicking on [End]. The second column is subtracted from the first one to obtain the duration of each case. If subtraction results in some non positive values, these are excluded from the analysis as invalid (missing) cases. It is possible to use data in date format, in which case the number of days between the two dates will be used as the duration data (see 18.104.22.168. Date Data).
Censored: This variable is selected to show whether the termination time of a case is not known or the termination event has occurred. If the value in the column is zero then it is assumed that the case has been censored, which means that it has been excluded from the study before the termination event has occurred. By default, non-zero values (usually one) indicate that the termination event has occurred for this case. If a censor variable is not selected, then it is assumed that termination event has occurred for all cases.
The Cox Regression procedure provides a facility to change the default value of 0 for censored cases in a dialogue that pops up just after the Variable Selection Dialogue. You should be aware that once a change is made here, it applies to all Survival Analysis procedures throughout the session, which do not have a facility to edit this value.
Factor: One categorical data column may be selected by clicking on [Factor], to produce a table for all or some of the subgroups defined by this variable. Selection of a factor column is compulsory for the Survival Comparison Statistics procedure. In Cox Regression, the [Factor] selection has a slightly different meaning. It is still used to perform analysis on a number of subgroups, but the nature of the analysis changes to what is called a stratified analysis. The maximum number of strata is limited to six. For further information see 9.4.4. Cox Regression.